Gait analysis using LLE

In motion based gait recognition method, it is very important to detect gait cycle and extract dynamic features in a gait cycle. Gait has the property of cycle and there are similarities within adjacent frames of gait sequence. LLE is a nonlinear dimension reduction method which can preserve local geometry, so LLE is a suitable tool for analyzing gait cycle and dynamic features in a gait cycle. Experiments results show, 1-d LLE embedding of gait sequence has almost the same information as that of 2-d LLE embedding and is sufficient for gait analysis. Zero crossing of 1-d or 2-d LLE embedding shows gait cycle and the shape of 1-d or 2-d LLE embedding shows the dynamic features of gait.